The data analysis and modeling group provides mathematical analysis of the kinetics of biodistribution of radiopharmaceutical as both a research and a service component. In particular, we will extend residue analysis to each of the projects over the next five years and apply novel heterogeneity analysis to the characterization of tumors using PET imaging. The clinical utility of image-based information needs to be tested in terms of what additional information the PET-derived measurements contribute to the prediction of patient outcome at the treatment planning stage. Each of our hypotheses describes an observational study with imaging data, assays of relevant biomarkers, and measures of patient response/survival. Our goal is to identify imaging procedures where a clinically important improvement in predictive accuracy, beyond that attained with existing clinical and diagnostic methods, can be achieved. Because the improved prediction will have clear therapeutic implications and our results will pave the way for these hypotheses to be tested in future cooperative clinical trials, the consultants in biostatistics play a critical role in focusing protocols and directing data analysis in anticipation of cooperative prospective trials. In order to increase the level of biostatistical involvement in protocol design and hypothesis testing, this new data analysis core will include a sub-contract with F O'Sullivan and the local leadership of M Muzi and D Mankoff plus new effort by L Kessler that are essential to the quality and overall productivity of our research program. It will build on our research to generalize the role of the more well established imaging agents through several smaller patient trials in a range of tumor histologies. Our goal is to develop a strategy to convince the cancer community in general and FDA in particular that mechanistically validated imaging agents can be used in parallel with new treatment trials. That is, qualification of an imaging agent should not be tumor-specific; it should be validated for a specific characteristic of the tumor phenotype.. Our goal is to make a convincing case that new imaging methods should be engaged in treatment trials sooner rather than later. The recent addition of L Kessler, a health policy expert who recently worked for FDA, to our faculty is important for this objective.